Help With VS1002 and Eagle

I’m having a lot of trouble designing a board for the VS1002 in EAGLE.

First I’m kind-of new to EAGLE (Did a few through-hole boards to learn the basics). When I run the Auto Router it never finishes (set to 8mils, the minimum for the PCB deal) Also when i run it through the EAGLE DRC I get width errors on all of the pins on the VS1002. I have an idea of what’s wrong but don’t know how to fix it nor do i know that SparkFun will accept it. If I route it to 5mils it STILL never finishes!??!?!?? I know that it could be routed it manually but I can barley hand route a through hole Board!!! :frowning: :cry: :x

LQFP DRC problem: what library part are you using for the vs1002? the one from cadsoft - vs10xx? The smd pads are .2 mm wide which is a tad less than 8mil (we’re talking tiny here) so you have to take the drc/sizes minimum width to 7.5 mil or so. This is not trace width which I would keep at 8 for signals. If the .2 mm pads don’t pass SFE DRC then you’ll have to edit the lib and make the pads bigger (.25mmx1mm). pretty easy - it took me about 5 minutes.

Routing: I really don’t like the eagle autorouter. It makes a hash of things. I route all my boards by hand. It may seem impossible at first but it’s really not that bad. I think the trace width is not your problem, the autorouter is just plain stupid. You can make the autorouter work but it takes a lot of effort to avoid stupid routes. Try putting a ground polygon on both the top and bottom layers (the entire board) and then routing power on the bottom first by hand. Then route short, direct and obvious traces and finally try the autorouter to finish it up (though at this point you might as well go all the way by hand).

You might also play with component rotation and placement to minimize the number of crossing airwires - that makes the job a lot easier (manual or auto). Don’t be afraid to rotate the parts in any orientation.

Yes, I’m using the VS10XX lib… LOL I made it. I thought all LQFP’s were the same so I used another part with the same package (LQFP-48 )

When you make a library how do you do all the measurement (length, width ect…)? Sorry for all the confusion for the people who used that Lib. :frowning:

But thanks for the helping this newbie into PCB designing :wink:

EDIT: I changed all the pads to .25mm x 1mm and with 8mils as the minimum on the DRC it passes :smiley:

TSPRAP:
Yes, I’m using the VS10XX lib… LOL I made it. I thought all LQFP’s were the same so I used another part with the same package (LQFP-48 )

When you make a library how do you do all the measurement (length, width ect…)? Sorry for all the confusion for the people who used that Lib. :frowning:

But thanks for the helping this newbie into PCB designing :wink:

EDIT: I changed all the pads to .25mm x 1mm and with 8mils as the minimum on the DRC it passes :smiley:

now that’s funny… small world

I usually start with the datasheet measurements. I’ve got a micrometer but measuring that small is a real challenge so I’ll just use it to verify that the chip is in the same ballpark as the datasheet. I’ve never see a datasheet be way off. Always switch to the “native” system (metric/imperial) that the datasheet uses and set the grid so that pins will properly line up. For example with an SOIC, I set grid to 25 mil and drop pins every other grid line. In the case of the lqfp-48, I used metric and set grid to .25mm when I looked at the package. It seemed to line up just fine (note, I haven’t routed with it so I can’t say 100%)…

I don’t think there was anything wrong with the package you started with, it’s just that the PCB shop might barf over the <8mil pads. But it never hurts to check.

One other point on that component - there are two chips that come with it. That’s a bit unusual. Usually, you would see the two chips as seperate lib entries. Nothing really wrong with the way you did it, just unusual.

Hope your routing goes well.

Phil